linux-kernel - Re: [PATCH] watchdog: qcom: fine tune the max timeout value calculation

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Thu, 11 Jan 2024 06:51:24 -0800
From: Guenter Roeck <linux@...ck-us.net>
To: Kathiravan Thirumoorthy <quic_kathirav@...cinc.com>,
 Bjorn Andersson <andersson@...nel.org>,
 Konrad Dybcio <konrad.dybcio@...aro.org>,
 Wim Van Sebroeck <wim@...ux-watchdog.org>
Cc: linux-arm-msm@...r.kernel.org, linux-watchdog@...r.kernel.org,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH] watchdog: qcom: fine tune the max timeout value
 calculation

On 1/11/24 01:49, Kathiravan Thirumoorthy wrote:
> To determine the max_timeout value, the below calculation is used.
> 
> 	max_timeout = 0x10000000 / clk_rate
> 
> cat /sys/devices/platform/soc@...017000.watchdog/watchdog/watchdog0/max_timeout
> 8388
> 
> However, this is not valid for all the platforms. IPQ SoCs starting from
> IPQ40xx and recent Snapdragron SoCs also has the bark and bite time field
> length of 20bits, which can hold max up to 32 seconds if the clk_rate is
> 32KHz.
> 
> If the user tries to configure the timeout more than 32s, then the value
> will be truncated and the actual value will not be reflected in the HW.
> 
> To avoid this, lets add a variable called max_tick_count in the device data,
> which defines max counter value of the WDT controller. Using this, max-timeout
> will be calculated in runtime for various WDT contorllers.
> 
> With this change, we get the proper max_timeout as below and restricts
> the user from configuring the timeout higher than this.
> 
> cat /sys/devices/platform/soc@...017000.watchdog/watchdog/watchdog0/max_timeout
> 32
> 
> Signed-off-by: Kathiravan Thirumoorthy <quic_kathirav@...cinc.com>

[ ... ]

> @@ -268,11 +271,11 @@ static int qcom_wdt_probe(struct platform_device *pdev)
>   		wdt->wdd.bootstatus = WDIOF_CARDRESET;
>   
>   	/*
> -	 * If 'timeout-sec' unspecified in devicetree, assume a 30 second
> -	 * default, unless the max timeout is less than 30 seconds, then use
> +	 * If 'timeout-sec' unspecified in devicetree, assume a 32 second
> +	 * default, unless the max timeout is less than 32 seconds, then use
>   	 * the max instead.
>   	 */
> -	wdt->wdd.timeout = min(wdt->wdd.max_timeout, 30U);
> +	wdt->wdd.timeout = min(wdt->wdd.max_timeout, 32U);

This is an unrelated change which would have to be discussed separately,
in a separate patch. But I really don't see the point of it.

Guenter